A few weeks back, Mike, Kent, and I spent a couple of days at Microsoft's annual systems seminar hearing about the future as Microsoft sees it. It should come as little surprise that MS sees OS/2 as part of everyone's future. Actually, I feel better about OS/2 and its chances of survivability after hearing some of MS's strategies. Nevertheless, as I've said before, if OS/2 is to be accepted, then MS needs to get out the 80386 version that will provide access to the 32-bit linear address space, allow multiple DOS sessions, include the high-performance file system, and more. But it isn't a trivial task to rewrite the several hundred thousand lines of assembly code that make up OS/2. Even though MS says this is one of their top priorities, it bothered me that we didn't see a technology demonstration of OS/2-386 at the seminar (MS is generally open about technology demos when they are making progress) and it makes me nervous about MS hitting their mid 1990 target for getting the 386 version out to market.
I was also interested in hearing more about what Microsoft intends on doing with object-oriented technology. When the words "C++ compiler" flashed on the overhead screen, I thought that we might be onto something. However, it turned out they weren't too specific on their plans because, it seemed, they had'nt really made up their mind yet. The internal debate seems to concern whether or not they should put out a stand-alone C++ compiler or add C++ extensions to existing C compilers. As best as I could tell, the current preference is to add extensions to the Microsoft C compiler sometime before the end of this year. But C++ isn't the only object-oriented path MS will be pursuing. They also talked about a "Visual" Basic which, among other things, will let people write graphic user-interface applications without having to do a lot of coding.
One thing that I found particularly interesting is that long-time language guru Greg Whitten, who has worked with all language implementations in the MS systems group over the past 10 years, recently moved over to the application development side of the company. His mission there seems to be two-fold: to direct the development of new object-oriented applications and to apply object-oriented technology to the applications development. According to Whitten, Microsoft's various applications make up several million lines of code and the object-oriented paradigm is one way for MS to keep things under control and on schedule. It's worth noting that the recent dramatic drop in MS earnings was attributed to the inability to get application products out in time, and they apparently believe 0OPs can prevent this in the future.
As if two days of OS/2 wasn't enough, the UniForum conference, which followed close on the heels of the MS seminar, left me with several impressions: 1. The Unix market is as fragmented as ever. 2. The widespread acceptance of the X Window protocol is one of the few consistent things that ran throughout the show; and 3. I'm tired of consortiums and "open" anything. Let's see ... there's the 88open consortium, the SPARC International consortium, the OSF consortium, and the Unix International consortium. Then there's Open Look, Open Systems, Open Desktop, Open Font, and X Open.
Here's an idea. Someone should go out and get the rights to the name Open Window because, considering the proliferation of windowing systems, some outfit is bound to come out with a product by this name and if you have the rights to the name, you might be able to make a couple of bucks. But then, somebody has probably already tied it up anyway.
Actually, the real future of the prefix "open" may be in vertical market applications. How about a package for dentists called Open Mouth, or a program for duck hunters called Open Season. Or maybe even a product for trouser manufacturers called Open Fly. Hey, will someone please open the door so I can get out of here.
Copyright © 1989, Dr. Dobb's Journal